In third generation (3G) Universal Mobile Telecommunications Systems (UMTS) and in particular in its evolved version System Architecture Evolution/Long Term Evolution (SAE/LTE), also referred to as Evolved Packet System (EPS), the concept of home base stations is introduced. In 3G, a home base station is referred to as a Home Node B (HNB) whereas in EPS it is referred to as a Home eNodeB (HeNB). A cell served by a HNB or a HeNB is commonly referred to as a femtocell. A home base station is assumed to be placed in a private home, utilizing the home owner's fixed broadband connection to access the core network. It is also assumed that the home owner handles the actual physical installation of the home base station. Hence, the deployment of home base stations cannot be planned, since it is largely outside the control of the operator. Another important property of the home base station concept is the potentially very large number of home base stations.
In the further description a 3G Home Node B is abbreviated HNB, an EPS/LTE Home eNodeB is abbreviated HeNB and the abbreviation HN is used to denote either a 3G Home Node B or an EPS/LTE Home eNodeB.
A HN connects to the operator's network via a secure tunnel (supposedly IPsec protected) to a security gateway at the border of the operator's network. Via this tunnel the HN connects to the core network nodes of the operator's core network. The 3GPP operator may also deploy a concentrator node in its core network between the HNs and the regular core network nodes. In the EPS standardization such a concentrator node is commonly referred to as a HeNB Gateway, which may be an optional node in EPS HeNB solutions. The corresponding node name in 3G UMTS standardization is HNB Gateway and this node is mandatory in 3G HNB systems. In this specification both HNB Gateways and HeNB Gateways may also be referred to as HN Gateways.
For both EPS and 3G UMTS the HN uses a broadband access network as part of the transport network. Through this setup a UE communicates via the HN and the core network like any other UE.
The HN concept is closely related to the concept of Closed Subscriber Group (CSG). A femtocell is also assumed to be a CSG cell. This means that only a selected group of subscribers are allowed to access the network through that cell. The CSG of a CSG cell is identified by a CSG ID, which is broadcast in the cell as a part of the system information. Typically each CSG cell has its own unique CSG ID but it is also possible to define the same CSG ID for multiple cells, thereby forming a CSG zone, in which the same selected group of subscribers is allowed access. Although a CSG cell in principle does not have to be a femtocell, the two terms are often used as synonyms.
Hence, all subscribers are not allowed to access a certain HN and a certain subscriber is not allowed to access all HNs. Under supervision of the operator the owner of a HN defines which subscribers are allowed to access a femtocell (CSG cell) of the HN (i.e. which subscribers that are included in the CSG of the femtocell). This is assumedly done through a web interface (or other “interface” between the HN owner and the operator), and the CSG data (or HN access list (which is an equivalent term assuming that the HN only serves one CSG)) is stored in a database in the operator's network. The HN owner would assumedly enter the allowed subscribers in the form of Integrated Services Digital Network (ISDN) numbers (Mobile Subscriber ISDN Numbers MSISDNs) or International Mobile Subscriber Identities (IMSIs) of the concerned subscribers.
FIG. 1 shows a mobile communication network with home base stations providing femto cells 110. The network comprises a core network (CN) 100. Base stations 102 providing macro cells 108 are connected to the core network via the 51 interface. The home base stations 106 are connected to the network via the home base station gateway 104 via the 51 interface. The X2 interface is a direct interface between base stations 102. Current standards do not specify that the HeNB GW 104 nor HeNBs 106 use the X2 interface, however potentially X2 may be used there as well.
One reason for having a base station gateway, exemplified by the HeNB GW between the home base stations exemplified by the HeNBs and the CN is to cope with the assumed number of HeNBs in a network which is very high (millions of HeNBs is one estimate).
This creates a possible scaling problem on the CN side as each HeNB will have its own S1 interface and it is assumed that the MMEs are not capable of handling millions of S1 interfaces.
The HeNB GW is needed to solve these scaling and signaling load issues basically to protect and offload the CN (MMEs).
The X2 interface between eNBs is mainly used for handover, so called ‘X2 initiated/based handover’. When X2 is set up between two eNBs, the identity of the eNB (eNB ID) and a list of information on the cells served by this eNB is sent to the other eNB and vice versa. Note that X2 is established only between eNBs that serve neighboring cells where handover may be performed or when Inter Cell Interference Coordination (ICIC) is needed. The X2 control plane messages (X2 Application Protocol, X2AP, messages) X2 SETUP REQUEST and X2 SETUP RESPONSE, are defined to establish an X2 interface.
The CN node that controls eNBs and HeNBs via the HeNB GWs is called Mobility Management Entity (MME). MMEs can be pooled in the sense that a group, or pool, of MMEs share the task of controlling a certain set of (H)eNBs, wherein the cells served by these (H)eNBs preferably cover a continuous area. There is a many-to-many relation between eNBs and MMEs, such that all MMEs in an MME pool are connected to all the eNBs that the MME pool controls. For HeNBs the many-to-many relation with the MMEs is preferably mediated, or proxied, by a HeNB GW.
The MMEs also handle control signaling towards the UEs that are connected via the (H)eNBs, using the Non-Access Stratum (NAS) protocols. For each UE that accesses the network via an eNB, the eNB selects an MME in the MME pool, which will handle the NAS signaling towards the UE for as long as the UE remains connected to the MME pool (i.e. as long as the UE remains connected to the CN via one of the (H)eNBs belonging to the MME pool). Since the eNBs mostly select different MMEs for different UEs, load-sharing is effectively introduced between the MMEs of an MME pool. When HeNBs are connected to the CN without an intermediate HeNB GW, the HeNBs act in the same manner as eNBs in this respect, i.e. they have S1 interfaces towards each MME in the MME pool and selects MME on a per UE basis. However, in the case where a HeNB GW is deployed, it is assumed that the MME pool support (e.g. selection of an MME) is performed by the HeNB GW on behalf of its connected HeNBs. That is, each HeNB only has a single S1 interface towards the HeNB GW, whereas the HeNB GW has an S1 interface to each of the MMEs in the MME pool. The HeNB GW must thus handle selection of MME per UE based on any information it receives in S1AP messages from the HeNBs. To keep the same MME for a certain UE the HeNB provides the available information received from the UE to the HeNB GW so that it can preferably again select the same MME as previously i.e. when still connected to the same MME pool.
An EPS network of more than moderate size will typically have several MME pools, where each MME pool controls a subset of the network's (H)eNBs. The X2 interface is only established between (H)eNBs belonging to the same MME pool. Between (H)eNBs belonging to different MME pools the handover signaling is conveyed via MMEs and the S1 and S10 interfaces where the S10 interface is the inter-MME pool (control plane) interface, i.e. an interface for control signaling between two MMEs belonging to different MME pools.
Within a PLMN each MME pool is uniquely identified by an MME Group Identity (MMEGI).
Identities and other configuration data in MMEs, eNBs and other nodes in a network are typically configured from a Operation, Maintenance, Administration and Provisioning (OAM&P) system. In this context it is often enough to refer to the O&M system, which can be seen as a subset of the OAM&P system. In the context of this document the terms OAM&P and O&M are used more or less as interchangeable equivalents.
When an eNB receives a measurement report from a UE which indicates that a cell belonging to another eNB may be a suitable handover target, the eNB may decide to establish an X2 interface towards the eNB serving the reported cell. The eNB will then request the UE to report the E-CGI (and some other parameters) of the reported cell (unless the UE has already done so). From the E-CGI of the reported cell the eNB then extracts the eNB ID of the eNB serving the reported cell (i.e. the eNB ID of the target eNB). The eNB then constructs an FQDN (Fully Qualified Domain Name) based on the eNB ID of the target eNB and includes this FQDN in a DNS query which it sends to a DNS server. The DNS server returns an IP address (or possibly a list of IP addresses) belonging to the target eNB and which can be used for X2 establishment. The eNB then initiates the X2 establishment by establishing a Stream Control Transmission Protocol (SCTP) connection and sending an X2 SETUP REQUEST message to the retrieved IP address (or one of the retrieved IP addresses in case the DNS server returned more than one IP address). The target eNB responds with an X2 SETUP RESPONSE message. A basic function of a DNS server is to perform the above described FQDN-to-IP address translation. A more general description of basic DNS server functionality is that it performs FQDN-to-Resource Record (RR) translation/resolution. There are several types of resource records. One of them, denoted “A record”, includes an IPv4 address, while another one, denoted “AAAA record”, includes an IPv6 address. Other types of resource records may include FQDNs, such as the “CNAME record” type or the “SRV record” type. Hence, in the description of the X2 establishment procedure above the DNS server returned an A record or an AAAA record (or a list of A or AAAA records (or even a combination of A and AAAA records) in case the DNS server returned multiple IP addresses). Note that resolution of an FQDN into an IP address via the DNS may involve a single or multiple DNS servers. In the latter case the multiple DNS servers are contacted recursively or iteratively (or a mix thereof). However, in order not to obscure the essentials of the invention the descriptions in this document are written as if a DNS query were always resolved by a single DNS server.
A mobile network may have several hundreds of thousands or millions of HeNBs. A capacity of the control nodes in the CN (Mobility Management Entities, MMEs) that will be able to handle that many HeNBs, i.e. that many control parts of S1 interfaces (S1-MMEs), is not reasonable. Therefore, one purpose of the HeNB GW, as illustrated in FIG. 1, is to conceal the large number of HeNBs from the CN. The HeNB GW will, from the CN (over the S1 interface), look like one eNB with many cells. The HeNB GW will act as an eNB proxy for all the HeNBs that are connected to the HeNB GW. From a HeNB, the HeNB GW will look like the CN (a regular S1 interface or possibly a slightly modified version thereof).